The higher the serum albumin, the better? Findings from the PRACTICE study

Aims The relation between hypoalbuminemia and coronary artery disease (CAD) has been established. However, the association of increased albumin level and outcomes of CAD has not been investigated. Methods There were 14 994 CAD patients from the PRACTICE study, which is a large, single center prospective cohort study based on case records and follow-up registry performed in the First Affiliated Hospital of Xinjiang Medical University from Dec. 2016 to Oct. 2021 in the present study. All the 14 994 patients were divided into five categories according albumin levels: <35 g/L group (n = 1 478), 35–40 g/L group (n = 5 007), 40–45 g/L group (n = 6 076), 45–50 g/L group (n = 1 835), and ≥50 g/L group (n = 598). Results A total of 448 all-cause deaths(ACD), 333 cardiac deaths (CD), 1 162 MACEs and 1 276 MACCEs were recorded during up to 60-months follow-up period. After adjusting for confounders, we observed a non-linear relation for either MACE or MACCE with the lowest risk at 45 g/L of albumin levels. A threshold value of albumin ≥50 g/L was associated with an increased risk for either MACE (adjusted HR=1.617, 95%CI:1.130–2.315, P = 0.009) or MACCE (adjusted HR= 1.439, 95%CI: 1.007–2.056, P = 0.045) in multivariable Cox regression model. For mortality, we only found decreased (<35 g/L) but not increased albumin level was associated with either ACD (HR=2.744, 95%CI: 1.631–4.617, P<0.001) or CD (HR=2.736, 95%CI: 1.484–5.045, P = 0.001). Conclusions In the present study, a U-shaped curve relation was identified between albumin levels and MACE and MACCE in CAD patients, with the lowest risk at 45 g/L levels.